Endgames Case Review

Refractory hypokalaemia in an elderly patient

BMJ 2017; 357 doi: https://doi.org/10.1136/bmj.j1575 (Published 12 April 2017) Cite this as: BMJ 2017;357:j1575
  1. Subramaniam Nagasayi, consultant in geriatric medicine1,
  2. Sandra Hui Li Yan, year 3 internal medicine medical officer1,
  3. Cleo Chiong, year 5, medical student2
  1. 1Department of Geriatric Medicine, Changi General Hospital, Singapore
  2. 2NUS Yong Loo Lin School of Medicine, Singapore
  1. Correspondence to N Subramaniam Subramaniam_nagasayi{at}cgh.com.sg

A 94 year old woman with mild dementia, peripheral vascular disease, and type 2 diabetes mellitus was discharged from hospital after completing antiviral treatment for a lower respiratory tract infection (influenza A3). She was readmitted four days later with secondary pneumonia and bacteraemia (meticillin sensitive Staphylococcus aureus, isolated from blood but not from lung or sputum samples). She was started on intravenous cloxacillin (2 g, every six hours) according to local guidelines. She had no episodes of diarrhoea, vomiting, respiratory muscle weakness, or paralytic ileus. On admission, her temperature was 38°C, blood pressure 141/70 mm Hg, respiratory rate 24 breaths/min, heart rate 114 beats/min, and peripheral capillary oxygen saturation 94%. Chest examination found normal heart sounds and right basal crepitations. Her abdomen was soft and non-tender; and her weight was 43 kg (body mass index 16 kg/m2).

By day five, her respiratory parameters had improved, but she had developed extreme fatigue with refractory hypokalaemia, despite oral and parenteral replacements (table 1).

Table 1

Serum levels and intravenous potassium replacement

View this table:

She continued on regular 100 mg aspirin and 20 mg omeprazole once daily and 250 mg tolbutamide twice daily. She was started on 600 mg slow release potassium chloride twice daily in addition to intravenous replacement.

Her haemoglobin level was 134 g/L (115-150 g/L), leucocyte count 13×109/L (4-10), C reactive protein 332 mg/L, urea 11.2 mmol/L (2.8-7.7), creatinine 95 umol/L (50-90), estimated glomerular filtration rate >60, sodium 130 mmol/L (135-145), potassium 4.4 mmol/L (3.5-5.3), glucose 20 mmol/L (3.1-7.8), and glycated haemoglobin 6.9%. QTc was 537 msc, and a chest radiograph showed right lower zone infiltrates.


  • 1. What electrolytes (along with potassium) would you monitor next in a patient with refractory hypokalaemia?

  • 2. If serum levels of other electrolyte levels are normal, how do you further investigate the cause of hypokalaemia in this patient?

  • 3. What is the most likely cause of this patient’s hypokalaemia?


1. What electrolytes (along with potassium) would you monitor next in a patient with refractory hypokalaemia?

Short answer

Serum magnesium levels needs to be checked in patients with refractory hypokalaemia to ensure adequate replacement.


Potassium secretion into the lumen by the cells of the connecting tubule and cortical collecting tubule is controlled by luminal potassium channels. This process is inhibited by intracellular magnesium and hence hypomagnesaemia and the associated reduction in intracellular magnesium concentration release. This inhibitory effect causes renal potassium wasting.1

2. If serum levels of other electrolyte levels are normal, how do you further investigate the cause of hypokalaemia in this patient?

Short answer

When the cause of hypokalaemia is unclear, first assess whether the potassium losses are caused by gastrointestinal or renal losses, by measuring spot urinary potassium excretion or potassium-creatinine ratio. Next, assess the acid-base status, which affects potassium homeostasis.


Common causes of hypokalaemia include2

  • Decreased potassium intake because the patient was given oral and parenteral supplements

  • Transcellular potassium shift caused commonly by insulin, drugs with increased β adrenergic activity, and increased blood cell production

  • Increased upper or lower gastrointestinal losses

  • Urinary potassium wasting caused commonly by diuretics, increased mineralocorticoid activity, and non-reabsorbable anions. Less common causes include renal tubular acidosis and salt wasting nephropathies.

If (as in this case), urinary potassium wasting is the likely cause, consider

Assessing urinary potassium excretion—The patient’s spot potassium excretion was 54 mmol/L (normal 5-15) and urine potassium-creatinine ratio was 16.8 (upper limit of normal: 1.5 mEq/mmol creatinine). A 24 hour urine collection is the most accurate method of assessment, but is impractical in an urgent setting. Alternatives are spot urinary potassium or urine potassium-creatinine ratio.3

Assessing acid-base status—assessment should consider:

  • Metabolic acidosis with low urinary potassium excretion is suggestive of lower gastrointestinal losses caused by laxative abuse or a villous adenoma

  • Metabolic acidosis with urinary potassium wasting is most often due to diabetic ketoacidosis or type 1 (distal) or type 2 (proximal) renal tubular acidosis

  • Metabolic alkalosis with urinary potassium wasting might be due to surreptitious vomiting

  • Metabolic alkalosis with urinary potassium wasting is either

    • normotensive—caused by diuretic use, vomiting, or Gitelman or Bartter syndrome. In this setting, measurement of the urine chloride concentration is often helpful (being normal in Gitelman or Bartter syndrome, high or low with diuretics), or

    • hypertensive—caused by surreptitious diuretic treatment with underlying hypertension, renovascular disease, or one of the causes of primary mineralocorticoid excess (Conn’s syndrome).

3. What is the most likely cause of this patient’s hypokalaemia?

Short answer

Cloxacillin. Without medications or the causes described, increased urinary losses of potassium are usually due to increased mineralocorticoid activity or increased distal delivery of sodium and water, particularly if they are accompanied by non-reabsorbable anions (penicillin) rather than chloride.


A calculated Naranjo probability score (obtained from a questionnaire for determining the likelihood that an adverse drug reaction is actually caused by the drug rather than other factors) indicated a “probable” causation by cloxacillin (a penicillin).4 Proton pump inhibitors are also known to cause hypokalaemia,5 although a score of 0 made this unlikely.

Penicillins are believed to cause hypokalaemia through two different mechanisms:

  • Penicillin acts as a non-absorbable anion in the cortical collecting duct. This results in increased lumen negative charge, an increase in sodium reabsorption, and thus potassium secretion.67 This is known to be a class effect across all penicillins.8

  • Cloxacillin sodium contains 2.1 mmol of sodium for every gram, and its administration leads to solute diuresis by increasing the flow rate in the cortical collecting duct and potassium excretion via the big potassium channels.910 Concurrent volume depletion increases the potassium secretion by increasing aldosterone secretion and decreasing distal chloride delivery. Aldosterone increases potassium secretion by increasing activity of the Na-K-ATPase pump and the luminal membrane permeability for potassium in principal cells. Most intravenous penicillins are salts of sodium; therefore these effects apply to other penicillins, such as cloxacillin, flucloxacillin, and piperacillin.

The effects described occur more in patients receiving high dose intravenous rather than oral penicillin.

Patients with low body mass index who are receiving high dose penicillin treatment are at risk of developing severe hypokalaemia. A reduction in dose or change of appropriate antibiotic is therefore necessary.

Patient outcome

On day nine, cloxacillin was switched to cefazolin 2 g three times daily for a further five days, and the hypokalaemia was corrected within 24 hours of the change. Severe fatigue and generalised muscle weakness delayed the patient’s recovery, however. Had the cause of hypokalaemia been identified and corrected earlier, her overall functional recovery might have occurred earlier than four weeks.


  • Competing interests: We have read and understood BMJ policy on declaration of interests and declare the following interests: None.

  • Provenance and peer review: Not commissioned; externally peer reviewed.

  • Patient consent obtained.


View Abstract

Log in

Log in through your institution


* For online subscription